Getting stills out of moving footage
Video is a sequence of complete images, so pulling one out is a matter of decoding to the right moment and saving what is there. No reconstruction is involved and nothing is approximated — an extracted frame is exactly the picture the video would have displayed at that instant, at the video's own resolution.
The interval control decides how many you get. One frame per second turns a two-minute clip into a 120-image contact sheet. Fifteen per second gets you close to every frame, which is what you want when the moment you are after lasts a fraction of a second.
Why there is a frame limit
Extraction stops at 300 images per run. The constraint is not decoding speed — FFmpeg will happily decode thousands — it is memory.
Every extracted frame has to be held in the browser tab as a complete image while the ZIP is assembled. Three hundred 1080p PNGs is well over a gigabyte of live allocation, and browsers do not fail gracefully when a tab exceeds what it can hold: they discard it, taking your work with it. Stopping at a number that reliably completes is better than attempting more and losing everything.
To cover a long video, use a longer interval. To capture a specific moment in fine detail, trim the clip first and then extract densely from the short result.
Why extracted frames sometimes look soft
This surprises people who expect a still from a 4K video to look like a 4K photograph. It generally will not, for two reasons that are both properties of the source rather than of the extraction.
First, motion blur. A video camera's shutter is open for a large fraction of each frame interval, so anything moving is smeared within the frame itself. That blur is recorded, not added later, and no amount of careful extraction removes it.
Second, compression. Video codecs spend most of their data on frames that resemble their neighbours and very little on frames during rapid change. A still grabbed from a fast pan was encoded with a small fraction of the information a static shot would have received.
The practical consequence: still frames from static or slow-moving footage look excellent, and frames from fast action look like what they are.
Common uses
- Grabbing a thumbnail or cover image from a video
- Finding the exact frame where something happened
- Building a contact sheet to see a long recording at a glance
- Pulling reference stills for illustration or analysis
- Extracting frames from a screen recording for documentation